Keystone Species: The Animals That Hold Ecosystems Together

One Starfish Changed How We Understand Ecosystems

In 1966, ecologist Robert Paine published a paper in the American Naturalist describing a simple but profound experiment on a rocky intertidal shore in Washington State. He manually removed all specimens of a single predatory sea star, Pisaster ochraceus, from an eight-meter stretch of shoreline and monitored what happened to the 15 other species living there. Within a year, mussels—previously kept in check by sea star predation—had expanded to cover the entire experimental area, crowding out barnacles, limpets, chitons, nudibranches, and algae. Species diversity collapsed from 15 to 1. When Paine removed the sea star from the system, the community didn't just change. It unraveled. He coined the term "keystone species" in a 1969 paper, borrowing the architectural metaphor of the keystone—the central stone in an arch that, if removed, causes the entire structure to fail.

The concept reshaped ecology. It also reshaped conservation.

What Makes a Species a Keystone?

A keystone species has a disproportionate effect on its ecosystem relative to its biomass or abundance. This is the defining characteristic: a species whose impact on ecosystem structure and function is far greater than would be predicted from its numbers alone. Keystones typically exert their influence through predation, engineering the physical structure of the environment, or providing resources that other species depend on critically. Remove them, and cascading effects alter species composition and ecosystem function across multiple trophic levels.

Not all predators are keystones. A predator that merely reduces the abundance of common prey species it consumes is important but not a keystone. The keystone designation applies specifically when the predator's removal allows a competitively dominant prey species to monopolize resources and exclude other species—the mussel-and-Pisaster example being the archetype. The concept also extends beyond predators to include ecosystem engineers and mutualists that provide critical services for many other species.

Sea Otters and the Kelp Forest

Sea otters (Enhydra lutris) were hunted nearly to extinction for their pelts by the 18th and 19th century maritime fur trade, with populations reduced from an estimated 150,000–300,000 to fewer than 2,000 individuals by 1911, when international protection was established. The decades following their near-extirpation provided an unintentional natural experiment in keystone species removal. Sea otters eat sea urchins; sea urchins eat kelp. Without sea otters, urchin populations exploded, devastating kelp forests across the North Pacific. Kelp forests are highly productive ecosystems supporting hundreds of species—rockfish, lingcod, harbor seals, bald eagles, and dozens of invertebrate species all depend on kelp structure. Where sea otters recovered, kelp forests recovered with them; where otters remained absent, urchin barrens persisted.

The contrast became a benchmark for understanding trophic cascades. A 1995 study estimated that the economic value of kelp forest ecosystem services—fisheries support, nutrient cycling, coastline protection—in sea otter territory exceeded the economic value of the sea urchin and abalone fisheries that sea otters suppressed. The keystone removes more value than it destroys, in ecosystem accounting terms.

The African Elephant as Ecosystem Engineer

African elephants (Loxodonta africana) are ecosystem engineers—species that modify the physical structure of their environment in ways that create habitat for other species. Elephants push over trees, creating forest gaps that allow light to reach understory plants and creating dead wood habitat for a cascade of invertebrates, cavity-nesting birds, and fungi. They dig water holes during dry seasons that persist as water sources for dozens of other species long after the elephants move on. They disperse seeds of large-fruited trees—plants that evolved large fruits specifically as elephant attractants—across enormous distances in their dung. Without elephants, seed dispersal for these trees collapses.

In savanna ecosystems, elephants maintain the mosaic of woodland and grassland habitat that supports high species diversity by preventing woodlands from closing over grasslands during wet decades. The removal of elephants from savanna—through poaching pressure that reduced populations by more than 60% across Africa in the 1970s and 1980s—triggers woodland thickening that reduces habitat diversity and the species that depend on open grassland.

Wolves and the Yellowstone Trophic Cascade Quantified

The reintroduction of gray wolves (Canis lupus) to Yellowstone National Park in 1995 has provided the most intensively studied trophic cascade in conservation history. Before wolf reintroduction, elk herds grazed riparian vegetation continuously, preventing willows, aspens, and cottonwoods from establishing. Paine's keystone concept, developed in the intertidal, was applied to a terrestrial system at scale:

• Wolf predation reduced elk numbers (direct effect)
• The "landscape of fear" changed elk foraging behavior—elk avoided open valley bottoms where wolves could hunt effectively (indirect behavioral effect)
• Riparian willows and aspens recovered in previously over-grazed valley bottoms
• Beaver colonies colonized recovering willow stands; their dams altered stream hydrology
• Songbird diversity increased in recovering riparian vegetation
• Stream banks stabilized as plant roots held soil; river channels narrowed and deepened
• Fish populations increased in improved stream habitat

The cascade from apex predator to river morphology became one of ecology's most celebrated examples and one of its most debated: subsequent research has questioned whether wolf effects alone can explain the observed vegetation changes, with elk harvest by hunters, drought cycles, and bison population changes also playing roles. The consensus remains that wolves triggered meaningful trophic effects, though the full cascade story is more complicated than its popular telling suggests.

Keystones by Category

Beavers: Hydro-Engineers of Temperate Ecosystems

Beavers (Castor canadensis in North America, C. fiber in Eurasia) are among the most consequential ecosystem engineers in temperate regions. Their dams transform stream reaches from fast-flowing channels into slow, wide, shallow wetlands—creating entirely different habitat that dramatically increases species diversity. Beaver ponds slow water flow, allowing sediment to settle, raising local water tables, and increasing the duration that surrounding areas remain wet through dry seasons. Beaver wetlands support amphibians, waterfowl, wading birds, fish, invertebrates, and riparian mammals at densities far exceeding comparable streamside habitat without beavers.

Beavers were extirpated from much of Europe by the 19th century; their reintroduction to Scotland, Germany, the Netherlands, and other European countries has been used to restore wetland ecosystem services including flood attenuation and water quality improvement—services that cost governments enormous sums to provide through engineering when natural infrastructure has been lost.

Identifying Keystones and Rewilding Applications

Identifying keystones in ecosystems that have not been experimentally manipulated is difficult. The "community importance" metric—the change in community structure per unit change in a species' biomass—can identify candidates from observational data, and species interaction networks from food web studies can flag highly connected nodes. But the definitive test remains some variant of removal: documenting what happens when the species is lost to overhunting, habitat loss, or deliberate extirpation, and whether ecosystem structure changes disproportionately.

This understanding drives the rewilding movement's emphasis on restoring lost apex predators and ecosystem engineers. The logic is that restoring keystone species may deliver ecosystem recovery that would otherwise require intensive and indefinitely continued management intervention—harnessing ecological processes to do conservation work at landscape scales that no human agency has the resources to replicate artificially.